Engineers trying to fix a flawed model of how the Carp River floods in bad storms manipulated figures to make the river’s flow seem slower at critical points, newly released documents say.
The figures were rejected by other experts before anything was done with them, said a city manager working on the file, Don Herweyer, but he wasn’t sure why they’d been worked up in the first place. “Maybe it was just a modelling exercise to get the effects of different changes,” he said. “I don’t know.”
The documents, released under orders from the province’s information commissioner after a three-year fight over private consultants’ proprietary information and legal concerns, shed light on the tortured process of fixing a major engineering mistake.
Flooding along the Carp River in western Kanata has been a schemozzle for the city and developers trying to build on land between Kanata and Stittsville, and farther north near Scotiabank Place. Construction plans had to be put on hold for a year in 2008 when city engineers Darlene Conway and Ted Cooper, defying their bosses, found a mistake in a detailed model of the Carp: It didn’t account for all the extra water expected to run into the river from new pavement on its banks.
At its heart, the problem was that flood model, a supremely complex set of formulas and figures that assesses how quickly water would run into the Carp, how quickly it would run out the north end into the Ottawa River, and where it would back up along the way. It affects how close new buildings can go to the river and how crossings — such as bridges for Hwy. 417 and Hazeldean and Maple Grove roads — need to be built so they don’t get swamped or washed away. The Carp’s headwaters are in Glen Cairn in south Kanata, which has flooded repeatedly in bad storms. If the math is wrong, water ends up in places it’s not supposed to be.
Certain figures the model needs are called “roughness coefficients,” which put a value on how running water is slowed by what it runs over. Engineers are supposed to pull them from tables generated by decades of work by other experts. A smooth metal culvert gets a low number; a weed-choked dirt channel gets a high one.
“It began with the Army Corps of Engineers, believe it or not,” said Herweyer, a city manager in charge of development reviews. “It’s a pretty standardized approach.” Put the wrong number in and you’re telling your model incorrect things about the watercourse you’re trying to describe, which could be just as bad as not telling it about millions of litres of water running in from nearby parking lots.
The city had a battalion of consultants trying to fix the Carp River model in 2008 and 2009 and the newly released documents show numerous exchanges among them and city staff over months, trying to work out how best to solve the model’s problems. One stands out, sent by Aecom’s Paul Frigon to Don Moss of Greenland International Consulting, explaining some tweaks they were still making just a few days before their work was made public in late January 2009.
“[W]e have adjusted friction coefficients for bridges and culverts to address sites where water levels are critical,” Frigon wrote on Jan. 19, 2009, then listed them:
“at 417, the coefficients at the upstream bridges for Palladium and Maple Grove have been increased to eliminate impacts to 417
“[downstream] of Richardson, the coefficient at the Richardson culvert has been adjusted to 0.4 to minimize downstream impacts
“This adjustment has increased water levels at several locations through the restoration reach by 0.1 to 0.5m over existing conditions — however, the reach is [nearby landowners’] territory and their final designs will reflect these increased levels — the increased levels do not impact existing infrastructure.”
In other words, the model was tweaked to say water slows down upstream of the Highway 417 bridges over the Carp. That meant the model would say water levels would be nearly two feet higher upstream, but that could be dealt with by working allowances into new construction in the development territory.
Moss and Frigon struggled with the consequences of the changes, according to the documents. Frigon acknowledged to Moss that the “real” number for the Richardson Side Road culvert was actually 0.013, which meant he was using a number 30 times higher. Ultimately, Moss sent the email on to the city, warning that the new coefficients would have to be “defended” at some point.
The effect, raising water levels by nearly two feet in certain storms, wouldn’t “impact existing infrastructure” in 2009, but it’s not clear from the documents what effect there might be on new 417 bridges being built now as part of the widening of the highway through Kanata.
It’s not clear from the documents what happened next. Herweyer swears unreleased emails show the city rejected Frigon’s figures. Three other experts shot them down before they got anywhere, he said.
“In the end, the numbers were not used,” he said. “Maybe that’s just another way of saying the numbers couldn’t be defended.”
Regardless, a few weeks after the email exchange, the city announced the model had been fixed, it had a handle on how the Carp River works, and development around the river was on track to resume. The new 417 bridges are being constructed by the provincial government based on a pledge from the city that water in the Carp won’t rise beyond a particular level, Herweyer said. If the river ever runs higher than that, it’ll be the city’s job to fix things upstream by widening the river or building bigger storm ponds.
“We’re in regular communication with the [Ministry of Transportation],” Herweyer said. “There’s monitoring throughout.”